A 1% ceiling does not sound dramatic until you sit with it for a minute. After years of mixed signals, Russia finally opened a formal crypto market on September 1. Banks started lining up trading desks, custody plans, and digital depositories. Then the draft prudential rules arrived and quietly asked a blunt question: how much of a lender’s own money can actually sit next to Bitcoin, foreign tokens, and every derivative that moves when those prices move? The proposed answer is one percent of capital. That is not a ban. It is a leash.
What The Draft 1% Cap Really Changes For Banks
I’ve found that market openings get all the headlines, while capital rules decide who shows up. This draft is that second chapter. It would introduce two maximum risk ratios, N31 for a single credit institution and N32 for a banking group on a consolidated basis. Both would sit at 1% of own funds. Not average over a quarter. Not a year-end snapshot. Every operating day.
Miss the line on six or more operating days inside any 30 consecutive operating days, and the supervisor can step in. That cadence matters. A sloppy week of client flows, a sudden price spike, or a mismatched hedge could turn a “small book” into a compliance event. In my view, that daily test will shape product design more than the marketing copy about a newly legal market.
A regulated market without a tight capital fence is just a press release. The fence is where the real policy lives.
The coverage is wider than spot coins sitting on a bank balance sheet. Loans, derivatives, bonds, repo trades, guarantees, credit lines, and other contracts whose payments or value depend on cryptocurrencies or foreign digital instruments can count. If the cash flows lean on crypto, the exposure can lean on the 1% cap. That is the part some desks will underestimate if they treat this as a simple inventory limit.
Why The Timing Feels Intentional
The market framework went live in early September. Nonqualified investors can buy eligible liquid cryptocurrencies up to 300,000 rubles a year through each intermediary after a suitability test. Qualified investors can trade without that purchase ceiling, though testing still applies. Weeks later, the same supervisor flagged cryptocurrencies and stablecoins as a financial market risk, including the worry that digital assets could act as substitutes for the ruble and that investors can lose everything.
So the sequence is not accidental. First, build the rails. Then cap how much bank capital can ride those rails. I think that is a familiar pattern in banking: permit the activity, then price it so expensive in capital terms that only controlled slices survive.
Group 1 And Group 2 Are Not Academic Labels
The draft splits crypto-linked transactions into two buckets. The split is not about slogans. It is about sanctions risk, settlement quality, and physical liquidity. That last phrase is easy to skip. It should not be skipped. If you cannot move or convert the position when you need to, the hedge on paper is not a hedge in a crunch.
Group 1 is the cleaner side of the ledger. It can include certain exchange-traded cash-settled derivatives, qualifying over-the-counter derivatives, and instruments with counterparties that meet set credit standards. Some miner-related transactions can qualify when income from digital asset sales supports the structure. Deliverable derivatives and selected loans, credit lines, guarantees, repo trades, and bonds can land here when settlement is available in rubles or in currencies of countries not treated as unfriendly.
For those lower-risk positions, banks may offset long and short exposures. Offsetting is not free. Maturity gaps get discounts. The draft starts at 5% and climbs as the gap widens. A mismatch of 37 months or more carries an 85% coefficient. In plain language: a long dated long and a short dated short do not cancel one-for-one. The residual still eats the 1% budget.
Group 2 is the heavier bucket. Direct investments in cryptocurrencies and foreign digital instruments sit here. So do loans settled only in those assets, some repo trades, derivatives that fail Group 1 tests, and leftover crypto-linked deals. Exposure is the larger of the long or short position in each asset. No full netting. That single design choice will push banks toward simpler books and away from fancy two-sided inventories that look hedged until the supervisor says they are not.
| Item | Group 1 | Group 2 |
| Typical assets | Qualifying cash-settled derivatives, stronger counterparties, some miner-linked deals | Direct crypto, foreign digital instruments, non-qualifying derivatives |
| Netting | Long/short offset allowed with maturity haircuts | Larger of long or short; no full offset |
| Settlement preference | Rubles or currencies of non-unfriendly countries | Often asset-settled or otherwise excluded from Group 1 |
| Effect on 1% cap | Can be managed more tightly if books are matched | Consumes the cap faster |
The 1,250% Risk Weight Is The Real Price Tag
Limits and risk weights are cousins, not twins. The 1% ratio is a hard concentration fence. Capital adequacy is a separate, expensive conversation. Under the proposal, a bank’s aggregate crypto exposure and certain client positions for which the institution assumes responsibility would carry a 1,250% risk weight.
If you have lived with Basel-style math, that number is familiar. It is the classic “deduct it in spirit” treatment: one unit of exposure can demand more than twelve units of risk-weighted assets. You can still do the business. You just pay for it as if the position were almost equity-like in fragility.
Client assets tell a more nuanced story. Where a digital depository is responsible for losses from seizure or sanctions-linked restrictions, those client holdings can enter the relevant risk calculation. Where the bank does not bear that responsibility, the positions can drop out of N31 and N32 and instead take a 50% risk weight for capital adequacy. That split will decide how custody is legally packaged. The commercial pitch may be “we hold your coins.” The capital file will ask who eats the sanctions freeze.
- Cryptocurrencies and foreign digital instruments would not count as collateral when banks calculate provisions for possible losses.
- Derivatives tied to crypto or foreign digital instruments would carry a 36% risk factor in the proposed framework.
- Reporting on covered turnover plus the N31 and N32 ratios is planned from January 2027.
- Official publication is expected in the fourth quarter of 2026, with effect 10 days after publication.
That last pair of dates is easy to treat as bureaucracy. It is not. Between late 2026 and early 2027, banks will need systems that can map every crypto-linked cash flow into Group 1 or Group 2, apply maturity haircuts, compute daily ratios, and produce forms that are still being drafted. Anyone who has implemented a new prudential report knows the ugly truth: the policy text is shorter than the data dictionary.
What “Every Operating Day” Means On A Real Desk
Imagine a bank that wants to offer qualified clients a thin market-making book, a small inventory of liquid coins, and a handful of listed cash-settled contracts. On Monday the book is quiet. On Tuesday a client lifts a large offer. On Wednesday the price jumps and the residual inventory inflates mark-to-market exposure. On Thursday a hedge with a different maturity only partly offsets. Friday’s close still sits inside 1%. The next week repeats with worse timing. Six bad days inside thirty is not a theoretical stress. It is a busy month.
Perhaps the most interesting aspect is behavioral. When a limit is daily and the penalty window is rolling, desks stop thinking in monthly averages. They start thinking in buffers. A 1% hard cap often becomes a 0.4% internal cap, because nobody wants day five of six. That self-imposed haircut is how prudential rules shrink markets without writing the word prohibition.
I’ve seen similar patterns in other high-volatility books. The public rule looks modest. The private operating rule is tighter, because operational risk, model lag, and weekend gaps all steal room. Crypto does not sleep on Sunday. Bank reporting still does.
How The Open Market And The Bank Cap Fit Together
Retail access is deliberately small. Three hundred thousand rubles a year per intermediary, after a test, is not a floodgate. Qualified flow is the piece that can grow. Banks and brokers were already building toward that layer. One major lender has targeted early December for trading, custody, settlement, and digital depository services. Another has tested crypto trading in a brokerage app with a limited qualified group and has said a broader rollout depends on the finished rulebook, with plans for its own digital depository.
Those build-outs still make sense. Custody fees, brokerage spreads, and depository services can earn money without parking a giant proprietary book. The 1% cap plus the 1,250% weight just tells you where the profit is supposed to sit: in agency and infrastructure, not in a bank-sized directional bet.
Let clients touch the asset. Do not let the bank become the asset.
– A useful way to read the draft
Is that too neat? A little. Banks can still take risk. They just cannot take much, and they cannot pretend a messy hedge is a clean hedge. Foreign digital instruments sit in the same tent as coins. That matters if a product is “not quite crypto” in marketing and “very much crypto” in cash-flow terms.
Sanctions, Settlement Currency, And The Unfriendly List
Group 1 treatment leans on settlement in rubles or in currencies of countries not classed as unfriendly. That is policy as plumbing. If you cannot settle cleanly, you do not get the nicer netting. If counterparties fail credit standards, you do not get the nicer bucket. If a miner deal cannot show income from digital asset sales under the stated conditions, it may slide into Group 2.
Client loss allocation around seizure and sanctions restrictions is the other hinge. When a digital depository accepts that loss path, the exposure can re-enter the bank’s risk math. When it does not, the 50% weight path can apply and the concentration ratios may ignore the position. Lawyers will spend more hours on that sentence than traders will spend on the 1% headline. Fair enough. Headlines do not recapitalize a freeze.
In my experience, rules like this also change product menus. Deliverable contracts become less attractive unless settlement currency qualifies. Pure coin-settled lending looks expensive. Cash-settled listed derivatives with decent counterparties look comparatively sane. That is not a moral ranking. It is a capital ranking.
What Banks Will Likely Offer First
If I had to sketch a practical rollout, it would not start with a giant proprietary warehouse. It would start with boring, billable pipes.
- Qualified brokerage access to eligible liquid coins, with tight inventory and rapid flattening.
- Custody and digital depository services structured so loss responsibility is explicit.
- Cash-settled derivatives that can qualify for Group 1 and be offset with modest maturity gaps.
- Limited credit products only where settlement currency and counterparty quality keep them out of the expensive bucket.
- Group-level monitoring so N32 does not get blown by a subsidiary the head office barely watches.
Sber-scale infrastructure talk and limited brokerage pilots already point that way. The remaining question is not whether a bank can click “buy Bitcoin.” It is whether the legal wrapper, the depository statute, and the daily risk engine agree on what that click created.
How This Compares With A Simple Ban
A ban is easy to explain and hard to police at the edges. A 1% cap with punitive risk weights is the opposite. It admits the asset class, then makes bank balance-sheet storage costly. Households and qualified clients can still meet the market through licensed intermediaries. The banking system does not become the warehouse.
That model has a logic if the supervisor’s fear is twofold: substitution away from the national currency, and sudden losses that leak into depositors. You can dislike the fear and still see the architecture. Keep the ruble at the center of settlement. Keep bank capital mostly away from token price risk. Keep the door open just wide enough that activity happens in licensed rooms instead of informal ones.
Does it work? Ask again in 2028, after a full year of N31 and N32 reporting. Drafts look tidy. Live markets invent basis risk, weekend gaps, stablecoin depegs, and client concentrations that no table fully captures. The six-day trigger will be the first real stress test of whether this is a flexible guardrail or a tripwire.
The Reporting Clock And Why Operations Teams Should Care
Publication in the fourth quarter of 2026, effectiveness ten days later, reporting from January 2027. That is a short runway if forms are still being designed. Data lineage is the unglamorous hero here. A derivative might sit in a trading system as “commodity-like.” A repo might sit in a securities system as “collateral upgrade.” A guarantee might sit in a credit system as “off-balance contingent.” The draft says all of them can be crypto risk if value depends on coins or foreign digital instruments.
So someone has to tag the dependency. Not the marketing name. The payment formula. If a bond coupon jumps when a token index jumps, welcome to the ratio. If a credit line can be drawn in a coin, welcome to the ratio. If a client position is only excluded when the bank truly does not eat sanctions-related loss, the legal memo has to match the risk report. I have a soft spot for that kind of unglamorous alignment. It is where regulation either becomes real or becomes theater.
Daily control loop in practice: Identify crypto-linked cash flows Sort Group 1 vs Group 2 Apply netting and maturity coefficients Compare to 1% of own funds Store the day count inside the 30-day window File turnover plus N31 / N32 when reporting starts
Investors Should Read The Cap As A Signal, Not A Price Target
This draft does not tell you where Bitcoin goes next week. It tells you how much balance-sheet oxygen Russian banks can give the asset class. Thin oxygen usually means thinner proprietary flow, more agency business, and more sensitivity to rules around who counts as qualified.
For a private investor, the useful questions are practical. Which intermediary holds the coins? Who is on the hook if assets are frozen? Is the product cash-settled or deliverable? Does the bank need to warehouse risk to give you a quote, or can it match you and step away? Those questions were always smart. Under a 1% daily fence, they become the business model.
I’ll say this in ordinary language. If a bank advertises deep crypto inventory and instant everything, either the book is tiny relative to capital, or the structure pushes risk into a depository and client wrapper. Ask which one it is. Asking is not cynicism. It is adult money.
A Few Loose Ends That Still Matter
Collateral treatment is easy to miss and hard to live with. If coins and foreign digital instruments cannot be recognized as collateral for provision calculations, a loan that looks over-secured in a pitch deck may look naked in the credit file. That will cool coin-backed lending faster than any speech about volatility.
The 36% risk factor on related derivatives is another quiet lever. It is not the 1,250% hammer, but it still charges the options and forwards that desks use to keep client flow moving. Hedging is allowed. Hedging is not free.
Group consolidation under N32 closes a familiar loophole. Park the book in a smaller sibling, keep the parent ratio pretty, hope nobody adds it up. Consolidated measurement exists because that hope is not a control framework.
And yes, miners get a cameo. Some transactions with miners can qualify for the better group when income from digital asset sales meets the stated conditions. That is a narrow door, not a welcome mat for every hash-rate story.
The Human Read On A Technical Draft
Technical drafts hide a personality. This one is cautious, slightly suspicious of substitution for the ruble, and willing to let licensed activity exist if bank capital stays mostly elsewhere. I do not need to cheer or boo that personality to describe it. After the September opening, a zero-limit stance would have looked like bait-and-switch. An uncapped stance would have looked careless. One percent is the compromise number that says “yes, but barely.”
Will banks lobby for a higher ratio later? Probably, if client demand is real and systems work. Will the supervisor tighten if ruble substitution fears grow? Also possible. Drafts are photographs. Markets are movies. The still frame we have now is a small cap, a harsh risk weight, a two-group split, a daily test, and a 2027 reporting start.
If you work at a lender, the next useful act is not another strategy offsite. It is a mapping exercise: every contract whose value depends on coins or foreign digital instruments, every client wrapper that assigns sanctions loss, every maturity mismatch that will eat netting relief. If you are a client, the next useful act is simpler. Ask who owns the freeze risk. Then decide if 1% of someone else’s capital is a deep enough pond for the product you were offered.
That is the story under the headline. Russia did not slam the door. It installed a narrow slot, measured in own funds, watched every operating day, and priced leftover curiosity at 1,250%. Whether that slot becomes a market or a museum piece will depend less on slogans and more on whether banks can live inside a number that small without pretending the number is larger than it is.
